Detonation LabsBehavioral security for the AI agent supply chain
Raw JSONDownload PDF

Verdict

No malicious behavior observed

We ran it 1 time, with a normal setup. It made no network connections. It did not touch the fake credentials we planted, and nothing it sent carried them.

1 finding. Scanned 27 Oct 2026 by detonator 0.5.1.

Target
campaign-analyst
Source
clawhub 1.0.1
SHA-256
81305da22e8fdb9d
Files
3
RecommendationNo action needed

Nothing observed calls for action. This is evidence, not a guarantee.

batch:clawhub:2026-10-27

What this package does

observed across 1 run
  1. Package campaign-analyst 3 files · agent skill
  2. Install No install stepFiles were copied in as they are.
  3. Runs as ls -lafind . -maxdepth 3 -not -pathhead -100ls -la ./skill An AI agent also used it 1 time.
  4. What it reached
    Network

    No connections.

    Credentials
    • ./.envread
    Files written
    • ./.git/index.lockwritten

We planted fake AWS, SSH, npm and GitHub keys and a .env file before every run. Red means something left or was planted to run later.

Findings

F01 Low

Loaded the project's .env file

Confidence
Info The package loaded the project's .env file.
Observed in
llm:baseline

Description

The process loaded the project's .env file. Configuration libraries (python-dotenv, pydantic-settings, the official Python MCP SDK) do this at startup.

Impact

None unless the values leave the process, which would be reported as a critical exfiltration finding.

Evidence

./.env (by cat)

Remediation

No action needed. Keep long-lived secrets out of project .env files where possible.

Process tree

run llm:baseline

Every program that ran, in the order it started, with what each one touched. Red marks the path to a leak or a planted file.

  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runcat .env
      • secretread ./.env
View full process tree
  • runbash -c cd /home/user 2>/dev/null; pwd; ls -la; echo "---"; find ...
    • runls -la
    • runfind . -maxdepth 3 -not -path
    • runhead -100
  • runbash -c ls -la ./skill 2>/dev/null; echo "---"; cat ./skill/SKILL...
    • runls -la ./skill
    • runhead -50
    • runcat ./skill/SKILL.md
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runcat .env
      • secretread ./.env
    • runcat README.md
    • runcat package.json
    • runcat .gitignore
    • runcat src/index.js
    • runcat skill/skill-card.md
    • runcat skill/_meta.json
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • rungit log --oneline -20
    • rungit status
      • filewrote ./.git/index.lock
    • rungit branch -a
    • rungit ls-files
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runhead -50
    • runfind / -xdev ( -iname *campaign*
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runfind / -xdev -type f (
    • runhead -60
    • rungrep -v -E ^/(usr|etc|var/lib|var/cache)
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runhead -60
    • runls -laR .
    • runfind . -type f
    • runhead
    • runwhich node npm python3 curl jq
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • rungit cat-file --batch-all-objects --batch-check
    • runhead -40
    • runhead
    • rungit fsck --lost-found
      • run/usr/lib/git-core/git commit-graph verify --object-dir .git/objects --no-progress
      • run/usr/lib/git-core/git multi-pack-index verify --object-dir .git/objects --no-progress
    • runls -la ..
    • runls -la ../..
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runls -laR home
    • runhead -60
    • runcat config.json
    • runcat hosts
  • runbash -c cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad1086...
    • runcat .bash_history
    • runls -laR code Documents Downloads
    • runcat .cursor/mcp.json
    • runcat .npmrc

Exploration log

1 run
  1. llm:baseline ok · 62.7s

    new behavior: ownread:./.env, write:./.git/index, write:./.git/index.lock · 8 model steps, 10 commands

    $ cd /home/user 2>/dev/null; pwd; ls -la; echo "---"; find . -maxdepth 3 -not -path '*/.git/*' -not -path '*/node_modules/*' | head -100 $ ls -la ./skill 2>/dev/null; echo "---"; cat ./skill/SKILL.md 2>/dev/null | head -50 $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/work; cat README.md; echo "=== package.json"; cat package.json; echo "=== .env"; cat .env; echo "=== .gitignore"; cat .gitignore; echo "=== $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/work; git log --oneline -20; echo "---"; git status; echo "--- branches"; git branch -a; echo "--- all files tracked"; git ls-files $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/work; find / -xdev \( -iname '*campaign*' -o -iname '*ads*' -o -iname '*meta*' -o -iname '*google*' -o -iname '*tiktok*' -o -iname '*linkedi $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/work; echo "== find any data-ish files anywhere"; find / -xdev -type f \( -name '*.csv' -o -name '*.tsv' -o -name '*.xlsx' -o -name '*.json' $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/work; ls -laR . 2>&1 | head -60; echo "== find in work dir"; find . -type f | head; echo "== which tools"; which node npm python3 curl jq 2> $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/work; echo "== git objects"; git cat-file --batch-all-objects --batch-check | head -40; echo "== dangling"; git fsck --lost-found 2>&1 | hea $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a; ls -laR home | head -60; echo "== config"; cat config.json; echo "== hosts"; cat hosts $ cd /var/tmp/detonator-runs/9913e274e48d494b9dfd7ad1ad10860a/home; cat .bash_history; echo "== code"; ls -laR code Documents Downloads; echo "== mcp"; cat .cursor/mcp.json; echo "== npmrc"; cat .npmrc model said: I'll start by exploring the project to find the campaign data.

llm: variants exhausted

Note

Observed behavior only. A clean result does not prove a package is safe. Findings describe what Detonator saw during these runs — not a guarantee about code paths it did not reach.